Overview of Dynamics 365 Virtual Agent for Customer Service April '19 release

Note

These release notes describe functionality that may not have been released yet. To see when this functionality is planned to release, please review What’s new and planned for AI for Customer Service. Delivery timelines and projected functionality may change or may not ship (see Microsoft policy).

Dynamics 365 Virtual Agent for Customer Service is a new cloud service launching in Public Preview in April 2019. Built on the lessons learned by supporting multiple worldwide deployments of customer service bots, the Dynamics 365 Virtual Agent for Customer Service puts the power of deploying solutions to problems through a virtual agent in the hands of the subject matter experts—your customer service team.  

Virtual Agent for Customer Service will be available in Public Preview in April 2019 and will be generally available in the second half of 2019.

The key features coming in the Public Preview include:

  • Quickly discover support topics to automate: Using the integrated Dynamics 365 Customer Service Insights dashboards, quickly see where your human agents are spending the most time, and which issues are most costly or trending. Get suggestions from artificial intelligence on which issues to automate to have the most impact on the system.

  • Automate support topics with a virtual agent: Easily create a virtual agent to help automatically answer and resolve common issues using a simple, no-code graphical interface. You can create conversations from scratch, or use rich, industry-specific templates to get your virtual agent ready in minutes.

  • Variables: You can extract entities from the conversation and use them to populate variables that can be used within or passed on to other dialogs.

  • Actions: Not only can the virtual agent chat with users to troubleshoot their problems, it can also take actions on their behalf. Using Microsoft Flow, a customer service manager can enable an action or access data in your back-end systems, leveraging hundreds of connectors to common services shipping as part of the Power Platform.

  • Work together with your human agents: When the virtual agent can’t handle a problem, or a user asks for a human agent, the virtual agent pulls in a human agent to help. Either use the integration with Dynamics 365 for Customer Service or integrate with your specific case management and agent messaging system.

  • AI-powered analytics for continuous improvement: You can always see how well the virtual agent is doing, which topics are handled, and which topics require human intervention. Virtual Agent for Customer Service continually provides analytics and recommends actions to improve the customer service experience, helping you free up time from your human agents to focus on more complex cases.

Beyond Public Preview, we will continue to add functionality. Here are some highlights of what to expect:

  • AI-driven dialogs: Virtual agent can have multi-turn conversations without using a rigid step-by-step script authored by a customer service manager. Given a set of valid solutions and conditions, Virtual Agent for Customer Service automatically uses the next best question to ask—guiding the user to the appropriate solution.

  • Support for additional languages: Virtual Agent for Customer Service will be able to understand multiple languages, including English, French, German, Spanish, and Italian.

  • Support for additional channels: The same Virtual Agent for Customer Service will connect to all Microsoft Bot Framework channels, including Teams, Skype, Facebook, Slack, and others, helping your customers regardless of how they choose to communicate with you.

  • Support for telephony: Telephone calls are one of the costliest forms of customer service. You’ll be able to let your agent handle calls coming in to your call center using Microsoft’s speech recognition and text-to-speech technology. If the virtual agent can’t help, it will redirect the call back to the appropriate call queue and pass along what it has learned so a human agent can pick up the conversation where the virtual agent left off.